# Parel Full Docs Context

Updated: 2026-04-27

This file is a compact, AI-readable reference for the public Parel docs. The MDX pages are the source of truth; this file mirrors stable concepts, endpoints, examples, and the generated model catalog.

## Product

Parel provides a single OpenAI-compatible API for open and hosted AI models, plus BYOM GPU deployments for custom Hugging Face models. The public API is served from `https://api.parel.cloud/v1`.

## Authentication

Use bearer-token auth with a Parel API key.

```bash
curl https://api.parel.cloud/v1/models \
  -H "Authorization: Bearer pk-dev-YOUR_KEY"
```

Key format: `pk-{env}-{key}`. Example: `pk-dev-YOUR_KEY`.

## CLI

Install options:

```bash
brew install parel-cloud/tap/parel
scoop bucket add parel-cloud https://github.com/parel-cloud/scoop-bucket
scoop install parel
curl -fsSL https://parel.cloud/install.sh | sh
iwr -useb https://parel.cloud/install.ps1 | iex
```

Common commands:

```bash
parel auth login
parel auth token pk-dev-YOUR_KEY
parel models list
parel chat run -m qwen-3.5-35b "Explain streaming responses"
parel proxy start --port 7878
parel claude-code init
parel tasks get task_123
parel usage summary --json
```

## OpenAI-Compatible Chat

```bash
curl https://api.parel.cloud/v1/chat/completions \
  -H "Authorization: Bearer pk-dev-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen-3.5-35b","messages":[{"role":"user","content":"Write a short launch checklist."}]}'
```

Python SDK:

```python
from openai import OpenAI

client = OpenAI(api_key="pk-dev-YOUR_KEY", base_url="https://api.parel.cloud/v1")
response = client.chat.completions.create(
    model="qwen-3.5-35b",
    messages=[{"role": "user", "content": "Merhaba!"}],
)
print(response.choices[0].message.content)
```

## Endpoints

| Area | Endpoint | Notes |
| --- | --- | --- |
| Models | `GET /v1/models` | List available models |
| Chat | `POST /v1/chat/completions` | OpenAI-compatible chat and multimodal vision |
| Completions | `POST /v1/completions` | Legacy text completion compatibility |
| Images | `POST /v1/images/generations` | Sync or async image generation |
| Video | `POST /v1/videos/generations` | Async video generation |
| Audio transcription | `POST /v1/audio/transcriptions` | Speech-to-text |
| Audio speech | `POST /v1/audio/speech` | Text-to-speech |
| Embeddings | `POST /v1/embeddings` | Vector embeddings |
| Reranking | `POST /v1/rerank` | Rank documents against a query |
| Moderation | `POST /v1/moderations` | Content moderation |
| Tasks | `GET /v1/tasks/{task_id}` | Poll async generation jobs |
| Compare | `POST /v1/compare` | Run one prompt against multiple models |
| Queue | `GET /v1/queue/status` | GPU queue visibility |
| Deployments | `POST /v1/deployments` | BYOM deployment creation |

## BYOM

BYOM deploys a Hugging Face model onto managed GPU capacity and exposes it through Parel routes. Create a deployment, poll until it is running, then call deployment chat.

```bash
curl -X POST https://api.parel.cloud/v1/deployments \
  -H "Authorization: Bearer pk-dev-YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"name":"demo","huggingface_id":"Qwen/Qwen3.5-1.5B","gpu_tier":"rtx3090_24gb"}'
```

## Claude Code

`parel claude-code init` writes an idempotent launcher into the user's shell profile. It configures Claude Code to use Parel through the Anthropic-compatible proxy route while keeping the normal Claude Code command available.

```bash
parel claude-code init
parel claude-code status
parel claude-code uninstall
```

## Errors

Parel returns OpenAI-style error envelopes:

```json
{
  "error": {
    "message": "Model not found",
    "type": "invalid_request_error",
    "param": "model",
    "code": "model_not_found"
  }
}
```

Common status codes: 400 invalid request, 401 invalid key, 402 insufficient balance, 403 forbidden, 404 not found, 429 rate limited, 502 provider error, 503 temporary capacity issue.

## Public Docs Map

- EN Docs home: https://docs.parel.cloud/
- TR Ana sayfa: https://docs.parel.cloud/tr/
- EN Quickstart: https://docs.parel.cloud/quickstart/
- TR Hızlı başlangıç: https://docs.parel.cloud/tr/quickstart/
- EN Authentication: https://docs.parel.cloud/authentication/
- TR Kimlik doğrulama: https://docs.parel.cloud/tr/authentication/
- EN OpenAI-compatible API: https://docs.parel.cloud/openai-compatible/
- TR OpenAI uyumlu API: https://docs.parel.cloud/tr/openai-compatible/
- EN CLI: https://docs.parel.cloud/cli/
- TR CLI: https://docs.parel.cloud/tr/cli/
- EN Local proxy: https://docs.parel.cloud/cli/proxy/
- TR Lokal proxy: https://docs.parel.cloud/tr/cli/proxy/
- EN CLI command reference: https://docs.parel.cloud/cli/commands/
- TR CLI komut referansı: https://docs.parel.cloud/tr/cli/commands/
- EN Claude Code: https://docs.parel.cloud/claude-code/
- TR Claude Code: https://docs.parel.cloud/tr/claude-code/
- EN Instant BYOM: https://docs.parel.cloud/byom/instant/
- TR Anında BYOM: https://docs.parel.cloud/tr/byom/instant/
- EN Image Models: https://docs.parel.cloud/byom/image-models/
- TR Görsel modeller: https://docs.parel.cloud/tr/byom/image-models/
- EN Dedicated GPU: https://docs.parel.cloud/byom/dedicated-gpu/
- TR Dedicated GPU: https://docs.parel.cloud/tr/byom/dedicated-gpu/
- EN Weight prefetch: https://docs.parel.cloud/byom/weight-prefetch/
- TR Ağırlık ön çekme: https://docs.parel.cloud/tr/byom/weight-prefetch/
- EN Chat API: https://docs.parel.cloud/api-reference/chat/
- TR Chat API: https://docs.parel.cloud/tr/api-reference/chat/
- EN Completions API: https://docs.parel.cloud/api-reference/completions/
- TR Completions API: https://docs.parel.cloud/tr/api-reference/completions/
- EN Images API: https://docs.parel.cloud/api-reference/images/
- TR Görsel API: https://docs.parel.cloud/tr/api-reference/images/
- EN Video API: https://docs.parel.cloud/api-reference/video/
- TR Video API: https://docs.parel.cloud/tr/api-reference/video/
- EN Audio API: https://docs.parel.cloud/api-reference/audio/
- TR Ses API: https://docs.parel.cloud/tr/api-reference/audio/
- EN Embeddings API: https://docs.parel.cloud/api-reference/embeddings/
- TR Embeddings API: https://docs.parel.cloud/tr/api-reference/embeddings/
- EN Reranking API: https://docs.parel.cloud/api-reference/reranking/
- TR Reranking API: https://docs.parel.cloud/tr/api-reference/reranking/
- EN Moderation API: https://docs.parel.cloud/api-reference/moderation/
- TR Moderasyon API: https://docs.parel.cloud/tr/api-reference/moderation/
- EN Async tasks: https://docs.parel.cloud/api-reference/tasks/
- TR Asenkron işler: https://docs.parel.cloud/tr/api-reference/tasks/
- EN Compare API: https://docs.parel.cloud/api-reference/compare/
- TR Karşılaştırma API: https://docs.parel.cloud/tr/api-reference/compare/
- EN Models: https://docs.parel.cloud/api-reference/models/
- TR Modeller: https://docs.parel.cloud/tr/api-reference/models/
- EN Account: https://docs.parel.cloud/account/account/
- TR Hesap: https://docs.parel.cloud/tr/account/account/
- EN API keys: https://docs.parel.cloud/account/api-keys/
- TR API anahtarları: https://docs.parel.cloud/tr/account/api-keys/
- EN Usage: https://docs.parel.cloud/account/usage/
- TR Kullanım: https://docs.parel.cloud/tr/account/usage/
- EN Billing: https://docs.parel.cloud/account/billing/
- TR Faturalama: https://docs.parel.cloud/tr/account/billing/
- EN Errors: https://docs.parel.cloud/account/errors/
- TR Hatalar: https://docs.parel.cloud/tr/account/errors/
- EN Rate limits: https://docs.parel.cloud/account/rate-limits/
- TR Limitler: https://docs.parel.cloud/tr/account/rate-limits/

## Model Catalog

Total: 62 models

### LLM (Text & Chat) (30 models)
Endpoint: POST /v1/chat/completions

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `qwen3-coder-plus` | Qwen 3 Coder Plus | Alibaba | $0.8 in / $2 out per 1M tokens | 131K | chat, tool_use, reasoning |
| `qwen3-max` | Qwen 3 Max | Alibaba | $1.6 in / $6.4 out per 1M tokens | 262K | chat, tool_use, vision, reasoning |
| `qwen3-vl-plus` | Qwen 3 VL Plus | Alibaba | $0.8 in / $2 out per 1M tokens | 131K | chat, vision, tool_use |
| `qwen3.5-flash` | Qwen 3.5 Flash | Alibaba | $0.05 in / $0.1 out per 1M tokens | 1M | chat, tool_use, vision |
| `qwen-3.5-cloud` | Qwen 3.5 Plus | Alibaba | $0.3 in / $0.6 out per 1M tokens | 131K | chat, tool_use, vision, reasoning |
| `qwen3.6-plus` | Qwen 3.6 Plus | Alibaba | $0.8 in / $2 out per 1M tokens | 1M | chat, tool_use, vision, reasoning |
| `qwen-turbo` | Qwen Turbo | Alibaba | $0.05 in / $0.1 out per 1M tokens | 131K | chat, tool_use |
| `qwen-vl-max` | Qwen VL Max | Alibaba | $2.4 in / $6.4 out per 1M tokens | 131K | chat, vision, tool_use |
| `claude-haiku` | Claude Haiku 4.5 | Anthropic | $0.8 in / $4 out per 1M tokens | 200K | chat, tool_use, vision |
| `claude-opus` | Claude Opus 4.6 | Anthropic | $15 in / $75 out per 1M tokens | 1M | chat, tool_use, vision, reasoning |
| `claude-sonnet-4` | Claude Sonnet 4 | Anthropic | $3 in / $15 out per 1M tokens | 200K | chat |
| `claude-sonnet` | Claude Sonnet 4.6 | Anthropic | $3 in / $15 out per 1M tokens | 1M | chat, tool_use, vision, reasoning |
| `deepseek-v3.2` | DeepSeek V3.2 | DeepSeek | $0.27 in / $1.1 out per 1M tokens | 131K | chat, reasoning |
| `deepseek-v3.2-azure` | DeepSeek V3.2 | DeepSeek | $0.58 in / $1.68 out per 1M tokens | 131K | chat, tool_use, code |
| `gemini-3-flash` | Gemini 3 Flash | Google | $0.15 in / $0.6 out per 1M tokens | 1M | chat, tool_use, vision |
| `gemma4-26b` | Gemma 4 26B | Google | $0.07 in / $0.15 out per 1M tokens | 262K | chat, vision, video, tool_use |
| `gemma4-31b` | Gemma 4 31B | Google | $0.15 in / $0.4 out per 1M tokens | 262K | chat, vision, video, tool_use, reasoning |
| `llama-4-maverick` | Llama 4 Maverick | Meta | $0.25 in / $1 out per 1M tokens | 131K | chat, vision |
| `llama-4-scout` | Llama 4 Scout | Meta | $0.17 in / $0.17 out per 1M tokens | 524K | chat, tool_use, vision |
| `mistral-large-3` | Mistral Large 3 | Mistral | $0.5 in / $1.5 out per 1M tokens | 131K | chat, tool_use, code |
| `mistral-small-azure` | Mistral Small | Mistral | $0.3 in / $1.2 out per 1M tokens | 131K | chat, tool_use |
| `kimi-k2.5` | Kimi K2.5 | Moonshot AI | $0.77 in / $3.44 out per 1M tokens | 262K | chat, tool_use, vision, reasoning |
| `kimi-k2.5-bedrock` | Kimi K2.5 (Bedrock) | Moonshot AI | $0.6 in / $2.5 out per 1M tokens | 256K | chat, reasoning |
| `nemotron-3-super-120b` | Nemotron 3 Super 120B | NVIDIA | $1.3 in / $2 out per 1M tokens | 131K | chat, reasoning, code, tool_use |
| `gpt-5.3` | GPT-5.3 | OpenAI | $2.5 in / $10 out per 1M tokens | 131K | chat, tool_use, vision, reasoning |
| `gpt-5.4` | GPT-5.4 | OpenAI | $2.5 in / $15 out per 1M tokens | 131K | chat, tool_use, vision, reasoning |
| `gpt-5.4-mini` | GPT-5.4 Mini | OpenAI | $0.75 in / $4.5 out per 1M tokens | 1M | chat, tool_use, vision |
| `gpt-5.4-nano` | GPT-5.4 Nano | OpenAI | $0.2 in / $1.25 out per 1M tokens | 131K | chat, tool_use |
| `grok-4.1-fast` | Grok 4.1 Fast | xAI | $0.4 in / $1 out per 1M tokens | 131K | chat, tool_use |
| `grok-4.1-reasoning` | Grok 4.1 Reasoning | xAI | $0.4 in / $1 out per 1M tokens | 131K | chat, tool_use, reasoning |

### Image Generation (10 models)
Endpoint: POST /v1/images/generations

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `wan2.7-image` | Wan 2.7 Image | Alibaba | $0.02 in / $0 out per 1M tokens | - | image_generation |
| `wan2.7-image-pro` | Wan 2.7 Image Pro | Alibaba | $0.04 in / $0 out per 1M tokens | - | image_generation |
| `flux-schnell` | FLUX Schnell | Black Forest Labs | $0.003 in / $0 out per 1M tokens | - | image_generation |
| `flux2-flex` | FLUX.2 Flex | Black Forest Labs | See pricing | - | image_generation, image_edit |
| `flux2-pro` | FLUX.2 Pro | Black Forest Labs | See pricing | - | image_generation, image_edit |
| `gemini-image` | Gemini 3 Pro Image | Google | $0.15 in / $0.6 out per 1M tokens | 1M | image_generation, vision |
| `gemini-3-flash-image` | Gemini 3.1 Flash Image | Google | $0.15 in / $0.6 out per 1M tokens | 1M | chat, vision, image_generation |
| `imagen-4` | Imagen 4 | Google | See pricing | - | image_generation |
| `gpt-image-1.5` | GPT Image 1.5 | OpenAI | $5 in / $40 out per 1M tokens | - | image_generation, image_edit |
| `recraft-v4` | Recraft V4 Pro | Recraft AI | See pricing | - | image_generation |

### Video Generation (10 models)
Endpoint: POST /v1/videos/generations

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `wan-2.6-i2v` | Wan 2.6 Image to Video | Alibaba | See pricing | - | video_generation, image_to_video |
| `wan-2.6-t2v` | Wan 2.6 Text to Video | Alibaba | See pricing | - | video_generation, text_to_video |
| `seedance-1.5-i2v` | Seedance 1.5 Pro Image to Video | ByteDance | See pricing | - | video_generation, image_to_video |
| `seedance-1.5-t2v` | Seedance 1.5 Pro Text to Video | ByteDance | See pricing | - | video_generation |
| `seedance-2-i2v` | Seedance 2.0 Image to Video | ByteDance | See pricing | 131K | video |
| `seedance-2-t2v` | Seedance 2.0 Text to Video | ByteDance | See pricing | 131K | video |
| `veo-3.1-i2v` | Veo 3.1 Image to Video | Google | See pricing | - | video_generation, image_to_video |
| `veo-3.1-t2v` | Veo 3.1 Text to Video | Google | See pricing | - | video_generation |
| `kling-3-i2v` | Kling 3 Image to Video | Kuaishou | See pricing | - | video_generation, image_to_video |
| `kling-3-t2v` | Kling 3 Text to Video | Kuaishou | See pricing | - | video_generation, text_to_video |

### Speech-to-Text (3 models)
Endpoint: POST /v1/audio/transcriptions

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `faster-whisper-v3` | Faster-Whisper V3 | Community | See pricing | - | stt, turkish |
| `whisper-large-v3` | Whisper Large V3 | OpenAI | See pricing | - | stt, turkish |
| `whisper-large-v3-turbo` | Whisper V3 Turbo | OpenAI | $0.006 in / $0 out per 1M tokens | - | stt, turkish |

### Text-to-Speech (3 models)
Endpoint: POST /v1/audio/speech

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `elevenlabs-multilingual` | ElevenLabs Turbo v3 | ElevenLabs | See pricing | - | tts, turkish, voice_clone, multilingual |
| `azure-tts` | Azure TTS | Microsoft | $0.6 in / $12 out per 1M tokens | - | text_to_speech |
| `chatterbox-multilingual` | Chatterbox Multilingual | Resemble AI | See pricing | - | tts, turkish, voice_clone |

### Embeddings (5 models)
Endpoint: POST /v1/embeddings

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `qwen-embedding-v3` | Qwen Embedding v3 | Alibaba | $0.007 in / $0 out per 1M tokens | 8K | embedding |
| `titan-embed-v2` | Amazon Titan Embed v2 | Amazon | $0.003 in / $0 out per 1M tokens | 8K | embedding |
| `gemini-embedding-2` | Gemini Embedding 2 | Google | $0.015 in / $0 out per 1M tokens | 8K | embedding |
| `text-embedding-3-large` | Text Embedding 3 Large | OpenAI | $0.13 in / $0 out per 1M tokens | 8K | embeddings |
| `text-embedding-3-small` | Text Embedding 3 Small | OpenAI | $0.02 in / $0 out per 1M tokens | 8K | embeddings |

### Music Generation (1 models)
Endpoint: POST /v1/audio/music

| Model ID | Display name | Vendor | Pricing | Context | Capabilities |
| --- | --- | --- | --- | --- | --- |
| `elevenlabs-music` | ElevenLabs Music | ElevenLabs | See pricing | 131K | music_generation |